Sigmodontinae incertae sedis. Delomys has variously stood as a genus (Avila-Pires, 1960c; Gyldenstolpe, 1932; Tate, 1932f) or merged under Thomasomys, with or without subgeneric division (Ellerman, 1941; Moojen, 1952; Osgood, 1933d). As remarked by Osgood (1933d), Thomas in part created Delomys to taxonomically underscore the geographic disjunction between thomasomyines in SE Brazil and the diverse radiation of typical Thomasomys in the N Andes (e.g., see Reig, 1986). Defining morphological traits consolidated and contrasted with types species of Thomasomys and Oryzomys by Voss (1993); chromosomal variation reported by Zanchin et al. (1992b) and Bonvicino and Geise (1995). While the morphological differentiation of Delomys is comparable to that of other sigmodontine genera, its cladistic stature with regard to Andean thomasomyines—and that of other SE Brazilian endemics like Juliomys, Rhagomys, Phaenomys, and Wilfredomys (oryzomyine-thomasomyine plesions sensu Voss, 1993)—must await broader-based phylogenetic inquiry. In one such preliminary view using cytochrome b sequence data, Delomys erratically formed a basal clade with various other tribes depending upon the analysis (Smith and Patton, 1999); those authors arranged the genus as one of several "additional unique lines."
Number of recognized species given as four (Moojen, 1952), three (Bonvicino and Geise, 1995; Reig, 1986), two (Ellerman, 1941; Gyldenstolpe, 1932; Hershkovitz, 1998; Voss, 1993), or one (Cabrera, 1961; Corbet and Hill, 1991; Honacki et al., 1982). The three recognized here integrates the revision of Voss (1993) and karyological results of Zanchin et al. (1992b) and Bonvicino et al. (1995). Avila-Pires (1960c) associated Calomys plebejus Winge, 1887, as another form of Delomys; Voss (1993) accepted the generic assignment but regarded plebejus as a nomen dubium, considering its equivalence to any of the living species as indeterminate.